Semi-supervised learning for biomedical image segmentation via forest oriented super pixels(voxels)

Lin Gu, Yinqiang Zheng, Ryoma Bise, Imari Sato, Nobuaki Imanishi, Sadakazu Aiso

研究成果: Conference contribution

8 引用 (Scopus)

抄録

In this paper, we focus on semi-supervised learning for biomedical image segmentation, so as to take advantage of huge unlabelled data. We observe that there usually exist some homogeneous connected areas of low confidence in biomedical images, which tend to confuse the classifier trained with limited labelled samples. To cope with this difficulty, we propose to construct forest oriented super pixels(voxels) to augment the standard random forest classifier, in which super pixels(voxels) are built upon the forest based code. Compared to the state-of-the-art, our proposed method shows superior segmentation performance on challenging 2D/3D biomedical images. The full implementation (based on Matlab) is available at https://github.com/lingucv/ssl_superpixels.

元の言語English
ホスト出版物のタイトルMedical Image Computing and Computer Assisted Intervention − MICCAI 2017 - 20th International Conference, Proceedings
編集者Maxime Descoteaux, Simon Duchesne, Alfred Franz, Pierre Jannin, D. Louis Collins, Lena Maier-Hein
出版者Springer Verlag
ページ702-710
ページ数9
ISBN(印刷物)9783319661810
DOI
出版物ステータスPublished - 2017 1 1
外部発表Yes
イベント20th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2017 - Quebec City, Canada
継続期間: 2017 9 112017 9 13

出版物シリーズ

名前Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
10433 LNCS
ISSN(印刷物)0302-9743
ISSN(電子版)1611-3349

Other

Other20th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2017
Canada
Quebec City
期間17/9/1117/9/13

Fingerprint

Supervised learning
Image segmentation
Classifiers
Pixels

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

これを引用

Gu, L., Zheng, Y., Bise, R., Sato, I., Imanishi, N., & Aiso, S. (2017). Semi-supervised learning for biomedical image segmentation via forest oriented super pixels(voxels). : M. Descoteaux, S. Duchesne, A. Franz, P. Jannin, D. L. Collins, & L. Maier-Hein (版), Medical Image Computing and Computer Assisted Intervention − MICCAI 2017 - 20th International Conference, Proceedings (pp. 702-710). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); 巻数 10433 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-319-66182-7_80

Semi-supervised learning for biomedical image segmentation via forest oriented super pixels(voxels). / Gu, Lin; Zheng, Yinqiang; Bise, Ryoma; Sato, Imari; Imanishi, Nobuaki; Aiso, Sadakazu.

Medical Image Computing and Computer Assisted Intervention − MICCAI 2017 - 20th International Conference, Proceedings. 版 / Maxime Descoteaux; Simon Duchesne; Alfred Franz; Pierre Jannin; D. Louis Collins; Lena Maier-Hein. Springer Verlag, 2017. p. 702-710 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); 巻 10433 LNCS).

研究成果: Conference contribution

Gu, L, Zheng, Y, Bise, R, Sato, I, Imanishi, N & Aiso, S 2017, Semi-supervised learning for biomedical image segmentation via forest oriented super pixels(voxels). : M Descoteaux, S Duchesne, A Franz, P Jannin, DL Collins & L Maier-Hein (版), Medical Image Computing and Computer Assisted Intervention − MICCAI 2017 - 20th International Conference, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 巻. 10433 LNCS, Springer Verlag, pp. 702-710, 20th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2017, Quebec City, Canada, 17/9/11. https://doi.org/10.1007/978-3-319-66182-7_80
Gu L, Zheng Y, Bise R, Sato I, Imanishi N, Aiso S. Semi-supervised learning for biomedical image segmentation via forest oriented super pixels(voxels). : Descoteaux M, Duchesne S, Franz A, Jannin P, Collins DL, Maier-Hein L, 編集者, Medical Image Computing and Computer Assisted Intervention − MICCAI 2017 - 20th International Conference, Proceedings. Springer Verlag. 2017. p. 702-710. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-66182-7_80
Gu, Lin ; Zheng, Yinqiang ; Bise, Ryoma ; Sato, Imari ; Imanishi, Nobuaki ; Aiso, Sadakazu. / Semi-supervised learning for biomedical image segmentation via forest oriented super pixels(voxels). Medical Image Computing and Computer Assisted Intervention − MICCAI 2017 - 20th International Conference, Proceedings. 編集者 / Maxime Descoteaux ; Simon Duchesne ; Alfred Franz ; Pierre Jannin ; D. Louis Collins ; Lena Maier-Hein. Springer Verlag, 2017. pp. 702-710 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{52daeecc1d99479b935046aadb00e7c8,
title = "Semi-supervised learning for biomedical image segmentation via forest oriented super pixels(voxels)",
abstract = "In this paper, we focus on semi-supervised learning for biomedical image segmentation, so as to take advantage of huge unlabelled data. We observe that there usually exist some homogeneous connected areas of low confidence in biomedical images, which tend to confuse the classifier trained with limited labelled samples. To cope with this difficulty, we propose to construct forest oriented super pixels(voxels) to augment the standard random forest classifier, in which super pixels(voxels) are built upon the forest based code. Compared to the state-of-the-art, our proposed method shows superior segmentation performance on challenging 2D/3D biomedical images. The full implementation (based on Matlab) is available at https://github.com/lingucv/ssl_superpixels.",
keywords = "Image segmentation, Random forest, Semi-supervised learning, Super pixels(voxels)",
author = "Lin Gu and Yinqiang Zheng and Ryoma Bise and Imari Sato and Nobuaki Imanishi and Sadakazu Aiso",
year = "2017",
month = "1",
day = "1",
doi = "10.1007/978-3-319-66182-7_80",
language = "English",
isbn = "9783319661810",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "702--710",
editor = "Maxime Descoteaux and Simon Duchesne and Alfred Franz and Pierre Jannin and Collins, {D. Louis} and Lena Maier-Hein",
booktitle = "Medical Image Computing and Computer Assisted Intervention − MICCAI 2017 - 20th International Conference, Proceedings",

}

TY - GEN

T1 - Semi-supervised learning for biomedical image segmentation via forest oriented super pixels(voxels)

AU - Gu, Lin

AU - Zheng, Yinqiang

AU - Bise, Ryoma

AU - Sato, Imari

AU - Imanishi, Nobuaki

AU - Aiso, Sadakazu

PY - 2017/1/1

Y1 - 2017/1/1

N2 - In this paper, we focus on semi-supervised learning for biomedical image segmentation, so as to take advantage of huge unlabelled data. We observe that there usually exist some homogeneous connected areas of low confidence in biomedical images, which tend to confuse the classifier trained with limited labelled samples. To cope with this difficulty, we propose to construct forest oriented super pixels(voxels) to augment the standard random forest classifier, in which super pixels(voxels) are built upon the forest based code. Compared to the state-of-the-art, our proposed method shows superior segmentation performance on challenging 2D/3D biomedical images. The full implementation (based on Matlab) is available at https://github.com/lingucv/ssl_superpixels.

AB - In this paper, we focus on semi-supervised learning for biomedical image segmentation, so as to take advantage of huge unlabelled data. We observe that there usually exist some homogeneous connected areas of low confidence in biomedical images, which tend to confuse the classifier trained with limited labelled samples. To cope with this difficulty, we propose to construct forest oriented super pixels(voxels) to augment the standard random forest classifier, in which super pixels(voxels) are built upon the forest based code. Compared to the state-of-the-art, our proposed method shows superior segmentation performance on challenging 2D/3D biomedical images. The full implementation (based on Matlab) is available at https://github.com/lingucv/ssl_superpixels.

KW - Image segmentation

KW - Random forest

KW - Semi-supervised learning

KW - Super pixels(voxels)

UR - http://www.scopus.com/inward/record.url?scp=85029384171&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85029384171&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-66182-7_80

DO - 10.1007/978-3-319-66182-7_80

M3 - Conference contribution

AN - SCOPUS:85029384171

SN - 9783319661810

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 702

EP - 710

BT - Medical Image Computing and Computer Assisted Intervention − MICCAI 2017 - 20th International Conference, Proceedings

A2 - Descoteaux, Maxime

A2 - Duchesne, Simon

A2 - Franz, Alfred

A2 - Jannin, Pierre

A2 - Collins, D. Louis

A2 - Maier-Hein, Lena

PB - Springer Verlag

ER -